重力梯度传感器数据的模拟及其辅助导航

Data simulation of gravity gradient sensor and its aided navigation

针对目前国内重力梯度数据难以获取的现状,研究重力梯度正演算法中矩形棱柱法来获取重力梯度数据,并将其应用到重力梯度辅助导航中。同时针对目前大多数匹配算法依赖于惯性导航系统(INS)的初始位置误差和在重力梯度变化不明显的区域定位精度不高的问题,提出重力梯度差异熵调节的概率神经网络算法作为匹配算法。在6个网格的初始误差下进行仿真实验,结果表明:该算法是正确的和有效的,匹配精度优于传统的概率神经网络匹配算法。

In view of current domestic situation that gravity gradient data is difficult to be obtained,rectangular prism method is used to get gravity gradient data in gravity gradient forward algorithm and the data is applied in gravity gradient aided navigation.Meanwhile,in order to solve problems that most matching algorithms depend on initial position error of inertial navigation system（INS） positioning precision is not high in gravity gradient change obscure region,a new matching algorithm based on probabilistic neural network（PNN） modulated by gravity gradient variance entropy is proposed.Simulation experiment is carried out in six grids initial error,the results show that the improved algorithm is correct and effective,and matching precision is superior to the traditional probabilistic neural network matching algorithm.